Show HN: An API to extract texts from images and PDF files

gnosis wrote: post #3

This looks nice except for having to depend on your servers as a middle man. Any chance you could release the code as Free or open source so that its users can use it standalone on their own machines?

That's not planned at the moment but if we wouldn't find a way to monetize it, we would do it for sure.

Re: Show HN: An API to extract texts from images and PDF files

#6

Post by rpedela » Thu, Jul 18, 2013, 3:15 PM UTC

I like the concept and it is a good start. Pulling text from PDFs is especially painful. I think the output format needs improvement. It is just a large array of strings. It seems like the strings are sometimes a single line, and sometimes not. My particular use case is extracting raw data from a PDF. I would like to see more structure to the output. For example, knowing where new lines, tabs, etc are located would be very helpful for parsing raw data.

Here is the PDF I used to test: https://www.gov.uk/government/uploads/system/uploads/attachm...

Is there a technical reason for the 1-2MB limit or is it arbitrary?

Re: Show HN: An API to extract texts from images and PDF files

#7

Post by rpedela » Thu, Jul 18, 2013, 3:17 PM UTC

Can I assume API keys are on the roadmap? I don't particularly like using my username and password.

Re: Show HN: An API to extract texts from images and PDF files

#8

Post by trez » Thu, Jul 18, 2013, 3:23 PM UTC

rpedela wrote: post #6

I like the concept and it is a good start. Pulling text from PDFs is especially painful. I think the output format needs improvement. It is just a large array of strings. It seems like the strings are sometimes a single line, and sometimes not. My particular use case is extracting raw data from a PDF. I would like to see more structure to the output. For example, knowing where new lines, tabs, etc are located would b…

Thanks for your comment!

That's something we can provide pretty easily and we would try to provide that in our next release. If you want us to help you with your specific problem, please send us an email at info@stamplin.com.

The limit has been set to prevent our server from crashing as we do not have, for the moment, the financial capability to support a massive server farm. Again, if this limit prevent you from using our API, we might move the limit up if you ask it by email.

Re: Show HN: An API to extract texts from images and PDF files

#9

Post by trez » Thu, Jul 18, 2013, 3:26 PM UTC

rpedela wrote: post #7

Can I assume API keys are on the roadmap? I don't particularly like using my username and password.

Yes, we'd like to increase security on each releases. It should be available in one of the next release.

Re: Show HN: An API to extract texts from images and PDF files

#10

Post by trez » Thu, Jul 18, 2013, 3:28 PM UTC

sebg wrote: post #4

smougel wrote: post #2

Any Feedback Welcomed

Looks good - does it do data tables? That's a big issue and something I've heard about (run into) many times...

Thanks for your comment! We would really appreciate if you could explain us in more details problems you faced. I am sending you an email if that's ok for you to discuss that.